Object Detection for Sign Language Recognition Using YOLOv8

N. Nagarani, S. Abinandhan, C. Elangovan, Pranesh Raja V. K.

2025

Abstract

Most people find difficulty in social communication with hearing-impaired people because they have less understanding of sign language. The hand gesture recognition technology translates the meanings of hand motions by detecting fundamental body signals which emerge through hand signals. The systems deliver necessary connectivity between educational institutions together with public service departments that operate within working environments. The YOLOv8 architecture creates an accurate system for hand gesture identification of English letters and numbers. YOLOv8, the latest version of the YOLO (You Only Look Once) series, combines real-time object identification with excellent computational efficiency and precision. The model structure includes three advanced segments that extract features from the backbone segment, then aggregates information through the neck before performing exact localization and classification in the detection head. The CSP (Cross Stage Partial) network backbone structure of the model minimizes operational costs without compromising its high feature performance capability. The implementation of PANet (Path Aggregation Network) in the neck segment establishes a superior method for feature movement until results are achieved. The detection head uses anchor-free predictions to achieve both high speed and accuracy in its output generation. The detection head provides precise results quickly because it employs anchor-free prediction systems. The detection head achieves rapid and accurate results by making predictions that do not depend on anchor points. Real-time hand movement detection functions as the core concept enabling this technology to perform effectively when users need it for social communication. Through its YOLOv8 technology, the system addresses communication barriers that enable hearing-challenged people to experience ordinary group conversations with public audiences without any obstacles.

Paper Citation

in Harvard Style

Nagarani N., Abinandhan S., Elangovan C. and K. P. (2025). Object Detection for Sign Language Recognition Using YOLOv8. In Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies - ICRDICCT`25; ISBN 978-989-758-777-1, SciTePress, pages 246-252. DOI: 10.5220/0013896100004919

in Bibtex Style

@conference{icrdicct`2525,
author={N. Nagarani and S. Abinandhan and C. Elangovan and Pranesh Raja K.},
title={Object Detection for Sign Language Recognition Using YOLOv8},
booktitle={Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies - ICRDICCT`25},
year={2025},
pages={246-252},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013896100004919},
isbn={978-989-758-777-1},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies - ICRDICCT`25
TI - Object Detection for Sign Language Recognition Using YOLOv8
SN - 978-989-758-777-1
AU - Nagarani N.
AU - Abinandhan S.
AU - Elangovan C.
AU - K. P.
PY - 2025
SP - 246
EP - 252
DO - 10.5220/0013896100004919
PB - SciTePress